Skip to main content

Parallel Compute

Description​

ParallelCompute is OpenCore's runtime API for handling CPU-intensive tasks without blocking the main GTA/FiveM thread. It uses a Worker Pool to execute TypeScript code in background threads.

This is critical for maintaining high server FPS when performing heavy operations like complex pathfinding, data processing, or large-scale mathematical simulations.

How it Works​

Worker Pool​

The service manages a pool of background workers (Native Workers). When you submit a task, it is queued and picked up by the first available worker.

Task Execution​

Tasks must be defined in a way that they can be serialized and sent to a worker. The service handles the communication and returns a Promise that resolves when the worker completes the task.

API Methods​

run()​

Executes a task in a background worker.

async run<T, R>(taskName: string, data: T): Promise<R>

getStats()​

Returns information about the current state of the pool: active workers, pending tasks in queue, and completed task count.

Why use Parallelism?​

FiveM runs the main script logic on a single thread. If you perform a calculation that takes 50ms, the entire server "freezes" for 50ms, causing lag and dropped frames.

By offloading that calculation:

  1. The main thread stays free to process networking and game logic.
  2. The calculation runs at full speed on a separate CPU core.
  3. You receive the result via a Promise when it's ready.

Example Use Cases​

  • Pathfinding — Calculating routes for complex AI.
  • JSON Processing — Parsing extremely large data structures.
  • Cryptography — Generating hashes or encrypted tokens.
  • Image/Data Compression — Processing assets on the fly.

Notes​

  • Isolation: Workers do not have access to the global scope of your resource. You must pass all required data via the task arguments.
  • Natives: FiveM Natives are generally not available inside worker threads. Parallelism is for pure TypeScript/JavaScript logic.
  • Overhead: Moving data between threads has a small cost. Only use this service for tasks that take significant time (>1ms).